EP0114517B1

EP0114517B1 - Mark position detecting method and apparatus

Info

Publication number: EP0114517B1
Application number: EP83307885A
Authority: EP
Inventors: Osamu Ikenaga
Original assignee: Toshiba Corp
Current assignee: Toshiba Corp
Priority date: 1982-12-27
Filing date: 1983-12-22
Publication date: 1987-03-11
Also published as: DD216530A5; US4701053A; JPS59119204A; EP0114517A1; DE3370200D1

Description

The present invention relates in general to a mark position detecting method and its apparatus. In particular, the present invention relates to a method of detecting a center position of a positioning reference mark formed on a mask for a pattern transfer by an energy beam, which is used for a mask defect inspection system, and an apparatus for executing the mark position detecting method.
In manufacturing the semiconductor integrated circuit (IC) chips, a reference pattern is formed on a mask for transferring a given micro-circuit pattern on a specimen. If the reference pattern contains a pattern defect such as disconnection, a manufacturing yield is reduced. To cope with this problem, a mask defect inspection system for inspecting a pattern defect prior to manufacturing the integrated circuits, has been developed. In this system, a mask member is illuminated to produce an optical signal representing a micro- circuit pattern formed on the mask member. This optical signal is compared with a theoretical data signal obtained on the basis of the design data used for forming the mask pattern. The optical signal representing a defect-contained mask pattern is not coincident with the theoretical data signal. The mask defect inspecting system detects the result of the comparison and judges whether or not a defect is present in the mask pattern.
For inspecting a pattern defect by using the mask defect inspecting system, it is essential to exactly align an optical system in this inspection system with the mask member in a high accuracy. An insufficient accuracy in the alignment between the optical system and the photomask results in an inaccuracy of the mask defect inspection. At the present stage, it is difficult to effectively and automatically detect the coordinates of a center position of a small positioning reference pattern previously formed on the mask.
It is therefore an object of the present invention to provide a new and improved mark position detecting method and its apparatus in which a center position of a positioning reference mark previously formed on a photomask may automatically be detected with a high precision.
According to one aspect of the present invention, there is provided an apparatus for detecting the position of a reference mark which is provided on a workpiece and has first and second edge portions disposed opposite to and substantially parallel with each other in a first direction and third and fourth edge portions disposed opposite to and substantially parallel with each other in a second direction normal to the first direction, comprising movable table means for supporting said workpiece, stationary light source means for radiating light on said workpiece positioned on said table means, and photodetector means for detecting an image of the edge portions of said reference mark produced by said workpiece transmitting light; characterised by edge position de- tecting means comprising means for moving said table means relative to said photodetector means so that the images of the two pairs of opposite edge portions are scanned by the photodetector means, the scanning for opposite edge portions being effected in opposite directions crossing the images of said edge portions, means for detecting the position of the table means during the movement, and means for producing data representing the position of each edge portion using the table position information and the signal of the photodetector means, and for computing the coordinates of the mark center based on the edge data.
According to another aspect of the present invention, there is provided a method for detecting a position of the reference mark which is provided on a workpiece and has first and second edge portions disposed opposite to and substantially parallel with each other in a first direction and third and fourth edge portions opposite to and substantially parallel with each other in a second direction normal to the first direction, said method comprising the steps of radiating light on said workpiece supported by a movable table, and detecting an optical image produced by said workpiece transmitting light, using a photodetector device, characterized by moving said table structure relative to said photodetector device so that the images of the two pairs of opposite edge portions are scanned by the photodetector device, the scanning for opposite edge portions being effected in opposite directions crossing the images of said edge portions, detecting the position of the table during the movement, producing data representing the position of each edge portion using the table position information and the signal of the photodetector device, and computing the coordinates of a mark center based on the data.
It should be noted that the aforementioned workpiece may be a mask for transferring a pattern using electron beam, light beam, or X-ray.
The present invention is best understood by reference to the accompanying drawings, in which:-

Figs. 1A and 1B are diagrams illustrating relative positions of a photodetector and a reference mark useful in explaining a conventional mark position detecting method;
Fig. 2 is a block diagram schematically illustrating a mark position detecting apparatus according to one embodiment of the present invention;
Fig. 3 is a diagram illustrating a photodetector formed of a plurality of photosensor elements provided in the Fig. 2 apparatus;
Fig. 4 shows a partial enlarged view illustrating a positioning mark previously formed on a photomask placed on the table of the mark position detecting apparatus according to the present invention;
Fig. 5 is a diagram illustrating relative positions of a photodetector and a reference mark useful in explaining a mark position detecting method according to the present invention;
Figs. 6A to 6D show a set of waveforms of detected signals at specific time points from all the photosensor elements of the photodetector of Fig. 3 in a Y-direction edge scanning step;
Figs. 7A to 7D are a set of characteristic diagrams illustrating time variations of the detected signals output from specific photosensor elements in the photodetector of Fig. 3 in the Y-direction edge scanning step;
Figs. 8A to 8D show a set of waveforms of detected signals at specific time points from all the photosensor elements of the photodetector of Fig. 3 in an X-direction edge scanning step;
Figs. 9A to 9D are a set of characteristic diagrams illustrating time variations of the detected signals output from specific photosensor elements in the photodetector of Fig. 3 in the X-direction edge scanning step; and
Figs. 10A and 10B cooperate to form a flowchart illustrating an execution flow of main steps in a mark position detecting method according to the present invention.

Before proceeding with the description of a preferred embodiment according to the present invention, a conventional mark position detecting technique will be described referring to Fig. 1. A photodetector 2 is provided for detecting a misalignment of an optical system in a photomask defect inspecting system with a rectangular positioning reference mark 4 formed on the photomask. The photodetector 2 is comprised of a plurality of photosensor elements arrayed in a line. By convention, a center position of the positioning reference mark 4 on the photomask is detected by the photodetector 2. The detected center position of the positioning reference mark 4 is then set at a proper target position to exactly align the optical system with the photomask.
For a linear mark edge 4a, a table (not shown) bearing a photomask thereon is moved so that the photodetector (signal-detector) 2 partially overlap with the mark edge 4a, while being disposed near the mark edge 4a. As shown in Fig. 1A, the mark edge 4a is one of the mask edges 4a and 4b of the positioning reference mark 4, which are disposed opposite to and parallel with each other and extend in a direction (X-direction in Fig. 1) orthogonal to a longitudinally extending direction of the photodetector 2 (Y-direction). When the photomask on the table is illuminated with light emitted from a light source (not shown), light transmitted through the photomask is incident on the detector 2. Then, the detector 2 generates electrical signals representing the information contained in the mask-transmitted light. More exactly, the plurality of photosensor elements sense the light, which contains the mask-transmitted light and has a specific distribution of incident light intensities. The output signals, or the output data, of the photosensor elements are compared with a predetermined threshold data. The output data compared are converted into digital signals, i.e. binary signals. The digitized signal contains a train of bits which are arrayed with two sections, one containing the successive bits with one logical state, while the other containing the successive bits with the other logical state. A point where the logical state changes in the bit train, i.e. the boundary between those sections, provides a first edge position data representing a first edge of the positioning reference mark 4, i.e. the mark edge 4a. A second edge, i.e. the edge 4b, disposed on the opposite side of and parallel with the mark edge 4a, is also position-detected in a similar manner to that for the mark edge 4a position detection. A center position of the positioning reference mark 4 as viewed in the Y-direction, i.e. the Y-coordinate of the mark center position, is computed on the basis of the first and second edge position data.
The table is then moved so that the photodetector 2 is positioned substantially parallel with and close to a mark edge 4c of the positioning reference mark 4. As shown in Fig. 1B, the mark edge 4c is one of the mark edges 4c and 4d linearly extending in the longitudinal extension (Y-direction) of the linear array of the photosensor elements of the photodetector 2. Following this table positioning, similarly, the mark placed on the table is illuminated with light. A specific photosensor element among the photosensor elements of the signal-detector 2 thus detects transmission light signal from the mask to generate an electrical detection data. The data is compared with a predetermined threshold data, and are converted into digital signal. This processing is repeated while continuously moving the table in a specific direction, i.e. X-direction, normal to the mark edge 4c. A third edge position data representing the position of the edge 4c is obtained by detecting a boundary between the detection data derived from the specific photosensor element and the threshold data. Further, the mask bearing table is continuously moved against the edge 4d opposite to and parallel with the third edge 4c, in the direction normal to the edge 4d. The mask-transmitted signal derived from the specific photosensor element is processed in a similar manner to the above one, and a fourth edge position is obtained. Then, the center position on the positioning reference mark 4 as viewed in the X-direction is computed using the third and fourth edge positions thus obtained. Finally, the center position, i.e. its X- and Y-coordinates, on the mark is obtained.
The conventional mask position detecting technique, however, has the following disadvantages. In case of detecting the position of the mark edge disposed in the direction perpendicular to the signal-detector 2, the mark edge detecting accuracy is limited due to the pitch of photosensor elements. Further, if the photosensor element, even single, is defective in the photodetector 2, the mark edge detecting accuracy is degraded depending on a size of the light sensing area of the defective element. Further, for detecting the mark edge position of the edge disposed parallel with the photodetector 2, the mark edge position is analyzed using the digital signal collected by the photodetector 2 in synchronism with the movement of the table. For this reason, an accuracy of the mark edge position detection is directly limited to by a noise signal contained in the photosensor signal. Furthermore, one of the paired edges is position-detected at a change point from a glass portion of the mark to a chromium portion. The other edge is position-detected at a change point from the chromium pattern to the glass portion. Therefore, the determination of the mark center position is inevitably accompanied by an appreciable error. In other words, the table moving directions for detecting the paired mark'edges disposed oppositely and parallel with each other, are the same. Therefore, the table moving speed causes an error in the determination of the edge position, leading to an inaccurate determination of the mark center position.
Turning now to Fig. 2, there is shown an overall arrangement of a mark position detecting apparatus which is a preferred embodiment of the present invention. A movable table structure 10 is made up of an X-table 10a linearly slidable in one direction, i.e. an X-direction, and a Y-table 10b linearly slidable in the other direction, i.e. a Y-direction, orthogonal to the X-direction. The X-table 1 Oa is mounted on the Y-table 10b. A photomask 12 is stably placed on the table structure 10 in a known manner. The photomask 12 has a substrate (not shown) made of transparent material with a good light-transmittance such as glass. A table driver 14 is connected to the X- and Y-tables 10a, 10b, and properly drives the tables 10a and 10b under control of a computer control section 16 including a computer. A laser interferometer 18 measures the actual moving distances of the tables 10a and 10b to generate a measuring signal 20. The measuring signal 20 is supplied to a position measurement section 22. The position measurement section 22 measures positions of the tables 10a and 10b, to generate position data 24. The position data 24 is temporarily stored in a data buffer memory 26 and supplied to the computer control section 16 at an appropriate time.
A known light source 30 is disposed above the table structure 10. Light 32 radiated from the light source 30 is focused by a first lens 34 and the focused light is used as a light beam for forming a spot on the photomask 12. Light 36 transmitted through the mask 12 passes through a second lens 38 disposed under the table structure 10 and is incident on a signal detecting section 40. The signal detecting section 40 is constructed of a photodetector 42 having a plurality of photosensor elements 44-1,..., 44-n, as shown in Fig. 3. The transmitted light 36 coming through the second lens 38 forms an image on the photodetector 42. The image covers the overall surface of the photodetector 42. Detection data 46 obtained by the photodetector 42 is temporarily stored in another buffer memory 48 and is supplied to the computer control section 16.
With the arrangement shown in Fig. 2, the X-and Y-tables 10a and 10b constituting the table structure 10 are respectively movable bidirectionally along the X- and the Y-axes under control of the computer control section 16. That is, the X-table 10 is linearly movable in both the forward and backward directions along the X-axis, while the Y-table 10b is linearly movable in the both directions along the Y-axis.
Description will be given how the mask position detecting apparatus thus arranged detects the mark position. To begin with, the photomask 12 is stably placed on the table structure 10. The mask 12 has a thin positioning reference mark 50 on a predetermined surface area. The positioning reference mark 50 is made of chromium and rectangular in shape. In the present embodiment, the mark 50 as a square thin plate is formed on a predetermined area of the mask 12, together with a microcircuit pattern (not shown).
The driver 14 drives the table structure 10 on which the photomask 12 is placed in such a manner that the optical axis of an optical system containing the light source 30 and the first lens 34 becomes coincident with the vicinity of the center of the mark 50 (rough positioning-alignment). After the positioning of the table structure 10, the light source 30 emits the light beam 32. The light 32 is focused by the lens 34 and applied to the positioning reference mark 50 on the photomask 12. At the same time, the Y-table 10b is driven in one direction, for example, a direction as indicated by Y1 in Fig. 5. Thus, the centre of the light spot, which is formed on the mask 12 by the light emitted from the light source 30, linearly moves toward a first edge 50a. During this movement, the photodetector 42 having the photosensor array in the signal detecting section 40 receives the light 36 passing the photomask 12 to generate a plurality of scanning each produced by a photosensor element, signals together representing the positioning of the first edge 50a. The scanning signals are analog signals varying with the actual amount of light incident upon the photosensor elements, and are sequentially stored in the buffer memory 48. During the course of the above operation, the laser interferometer 18 supplies the position measurement section 22 with measuring signals 20 representing the present position of the Y-table in the Y1-direction. On the basis of the measuring signals 20, the position measurement section 22 computes an exact position of the Y-table 10b for each scanning signal 46 stored, and stores the result of the computation as a position data 24 in the data buffer memory 26.
In the above mark edge detecting operation, the photodetector 42 of the signal detecting section 40 produces scanning signals as shown in Figs. 6A-6D. Figs. 6A-6D schematically illustrate levels of the signals produced by all the photosensor elements 44 at time points tl-t4 during a period where the image of the mark edge 50a passes the photodetector 42. As seen from these figures, the signal level increases as the overlapping area of the photodetector 42 with the mark 50 decreases. The data stored in memories 26 and 48 are supplied to the computer control section 16 at an appropriate time. For obtaining a signal level variation with time at the individual photosensor elements 44-1,..., 44-n of the photodetector 42 using the scanning signals stored in the buffer memory 48, the computer control section 16 analyzes the position data stored in the data buffer memory 26 and the output signal levels derived from the individual sensor elements respectively corresponding to the position data. As a result, signal level variation with time of the output signal (or scanning signal), which is generated from each photosensor element in the photodetector 42, may be obtained as illustrated in Figs. 7A to 7D. The computer control section 16 computes the edge position of the mark edge 50a based on the change of the scanning signals. The computation of the edge position is performed on each of the time variations of the scanning signals derived from the sensor elements as illustrated in Figs. 7A-7D. A detected signal level variation of one sensor element 44-1, for example, is illustrated in Fig. 7A. In Fig. 7A, an envelope waveform 54 enveloping the peak points of the individual signals is obtained. A crossing point at which the envelope waveform curve crosses a predetermined threshold level is determined as a mark edge position. All edge position data thus obtained from the all sensor elements 44 are subjected to the statistical analysis (averaging process, for example) to finally determine the mark edge position with high precision.
The next step is to detect a position of the edge portion 50b which is on the other side of the edge. portion 50a seen in the Y-direction. The Y-table 10b is driven, by the table driver 14, to move in an opposite direction Y2 to that of the above movement. Subsequently, the signal detecting section 40 is driven. Then, the photodetector 42 of the signal detector 40 generates the scanning signals on the portion of the photomask 12 near the second edge 50b extending parallel to the first edge 50a. The scanning signals are stored in the buffer memory 48. At this time, the position data on the Y-table 10b generated when the scanning signals are obtained are stored in the data buffer memory 26 by the position measurement section 22 in the same manner as mentioned above. On the basis of those data, the actual position of the second mark edge 50b is computed by the computer control section 16 in the same manner as mentioned above. Finally, the position of the center P of the positioning reference mark 50, in the Y-direction i.e. the Y-coordinate of the center P, is computed.
The Y-table 10b is moved again to an initial position to allow the optical axis of the optical system in this apparatus to be substantially coincident with the center of the mark 50. Then, the X-table 10a is started to move in the direction of X1 in Fig. 5. In this embodiment, at this time, the photosensor elements 44-1, ..., 44-n constituting the photodetector 42, which are linearly arrayed in the X-direction, are successively crossed by the image of the third edge 50c in the direction of X2. Figs. 8A to 8D illustrate respectively the detected signals output from all the sensors 44 of the photodetector 42 at some time points t5 to t8, for example. The time sequence t5 to t8 does not mean that it directly follows the time sequence t1 to t4 in Fig. 6. In those figures, the signal levels are arrayed from left to right corresponding to those signals generated by the photosensor elements 44-1,..., 44-n, respectively. As can be seen the intensity of the light incident on the sensors located under the positioning reference mark 50, viz. located on the right in the figures, is weaker than that incident on the sensors located outside the positioning reference mark 50. The scanning data are stored into the buffer memory 48. The table position data obtained by the position measurement section 22 corresponding to the scanning data are stored in the other buffer memory 26. Time variation of the output signal levels of each of the sensor elements 44 can be illustrated as shown in Figs. 9A to 9D on the basis of the analysis made as in case of the Y-direction scanning. The position of the third edge 50c of the mark 50 is computed using an envelope wave 56 directly representing a signal level variation of each sensor element. It should be noted that, in this case, measurement data representing the mark edge position, which are determined on the basis of the output signals from the photosensing elements, include edge positions shifted from the center of the photodetector 42 by deviation values, which are n times (n = +1, +2,...) larger than or smaller than the alignment pitch of the photosensor elements. However, the corresponding positive and negative deviation values in one direction are cancelling each other, when computing a mean value of the measured position data. As a result, there can be obtained a mark edge position data accurately corresponding to the center of the photodetector 42. Thereafter, the X-table 10a is moved in the opposite direction X2 to that of X1. In an operation corresponding to the one mentioned above, the position of the fourth edge 50d is obtained. Finally, the position of the mark center P in the X-direction, i.e. the X-coordinate Xp of the mark center P, is computed using the third and fourth edge data. In this way, the coordinate values (Xp, Yp) of the mark center P are exactly obtained. Finally, the mark center position thus obtained is made coincident with the optical axis of the optical system by driving the table 10. At this point, the positioning work of the mask 12 is completed.
According to the mark position detecting technique of the present invention, it is possible to extremely accurately detect the center position of the rectangular mark 50 by using the photodetector 42 with a linear sensor array. A reason for this is that, in detecting the positions of the mark edges in both the X- and Y-directions, the level changing points corresponding to the edge positions can elaborately be determined on the basis of multi-level or analog change (this is made clear by the envelope curves 54 or 56) of the detected scanning signals from a plurality of sensor elements. Therefore, the present invention can completely solve the problem inherent to the prior art using the digital detected signal in binary form, in which when the sensor element, even single, is defective, the edge position detecting accuracy is degraded. According to the present invention, even when some of the photosensor elements are defective, the detecting accuracy is little degraded, to provide a stable mark position detecting operation.
Another useful feature of the present invention follows. In detecting two opposite edges of the rectangular mark 50, for example, the edges 50a and 50b, the mark 50 formed on the photomask 12 is moved in opposite directions so that these edges are scanned by the light 32 emitted from the light source 30. Such motion of the mark is realized by driving the X- and Y-table 10a and 10b in opposite directions during the scanning of the two mark edges. For example, in the step to obtain the Y- coordinate of the mark center P by detecting the positions of the first and second edges 50a and 50b, the positioning reference mark 50 is linearly moved in the Y1 and after that in the Y2 direction (or vice versa), which are opposite to each other. Therefore, the centre of the light spot is relatively moved from the chromium mark pattern 50 to the glass substrate, when both the mark edges 50a and 50b are detected. With the feature of the scanning iso- tropy, the error due to the anisotropic scanning can be eliminated, to improve the mark edge detecting accuracy. Even if the detected position data of the edges 50a and 50b contain errors, these are compensated by the computing process on the mark centre P.
As described above, in the mark position detecting method according to the present invention, the analog signals of the individual sensors linearly arrayed in the photodetector are collected in synchronism with the table positions. Therefore, the mark edge positions can be detected with a high accuracy without adverse influences from the sensor pitch and noise signal components. Moreover, in the method according to the present invention, for determining the first and second edge positions of the mark, the directions of the table motion are opposite to each other. This feature eliminates the anisotropic scanning in the X- and Y-directions. Further, the analog signals from the sensor array are analyzed corresponding to the table positions. This feature eliminates the need for the correction by the table motion speed. The data collecting direction of the photodetector is either of the directions from the glass portion to the chromium portion and the chromium portion to the glass portion. Therefore, errors due to the scanning operation in one direction are compensated by the scanning operation in the opposite direction. Furthermore, even when a defective sensor element is contained in the sensor array, the edge portion detecting accuracy is negligibly degraded by the presence of the defective element.
Although the present invention has been shown and described with respect to a particular embodiment, various changes and modifications are possible.
For example, the mark used is not limited to the rectangular shape, but may be an L-shaped one. Further, the transmitted light from the mark may be substituted by the reflecting light from the mark for obtaining the mark edge position data. The present invention is applicable not only for the mask defect inspecting apparatus, but also for other apparatuses in use of the mask positioning.

Claims

1. An apparatus for detecting the position of a reference mark (50) which is provided on a workpiece (12) and has first and second edge portions disposed opposite to and substantially parallel with each other in a first direction and third and fourth edge portions disposed opposite to and substantially parallel with each other in a second direction normal to the first direction, comprising movable table means (10) for supporting said workpiece (12), stationary light source means (30) for radiating light (32) on said workpiece (12) positioned on said table means (10), and photodetector means (42) for detecting an image of the edge portions of said reference mark (50) produced by said workpiece (12) transmitting light; characterised by edge position detecting means (14, 16, 18, 22, 40) comprising means (14) for moving said table means (10) relative to said photodetector means (42) so that the images of the two pairs of opposite edge portions are scanned by the photodetector means (42), the scanning for opposite edge portions being effected in opposite directions crossing the images of said edge portions, means (18, 22) for detecting the position of the table means (10) during the movement, and means (16) for producing data representing the position of each edge portion using the table position information and the signal of the photodetector means (42), and for computing the coordinates of the mark center (P) based on the edge data.

2. An apparatus according to claim 1, characterized in that said photodetector means (42) includes a plurality of photosensor elements (44) and means (40, 48) are provided generating from the signal of each of said photosensor elements (44) a multi-level scanning signal quantitized at predetermined time intervals, which signal serves to compute one of the first to fourth edge data.

3. An apparatus according to claim 2, characterized in that said computer means (16), is connected to said photodetector means (42) for receiving a plurality of scanning signals (46) with signal levels varying with time according to the light component passing a mark edge portion during the movement of the table means (10) for computing the envelope waveform (54, 56) of the signal levels of each of said scanning signals (46), and for computing the position of said mark edge portion at a detection error smaller than the pitch in the array of said photosensor elements (44) on the basis of a combination of each of said envelope waveforms (54, 56).

4. An apparatus according to claim 3, characterized in that said photosensor elements (44) are linearly arrayed in one of the table movement directions.

5. An apparatus according to claim 3, characterized in that said photosensor elements (44) are linearly arrayed in the first direction or in the second direction.

6. A method for detecting a position of the reference mark (50) which is provided on a workpiece (12) and has first and second edge portions disposed opposite to and substantially parallel with each other in a first direction and third and fourth edge portions opposite to and substantially parallel with each other in a second direction normal to the first direction, said method comprising the steps of radiating light (32) on said workpiece (12) supported by a movable table (10), and detecting an optical image produced by said workpiece (12) transmitting light, using a photodetector device (42), characterized by moving said table structure (10) relative to said photodetector device (42) so that the images of the two pairs of opposite edge portions are scanned by the photodetector device (42), the scanning for opposite edge portions being effected in opposite directions crossing the images of said edge portions, detecting the position of the table (10) during the movement, producing data representing the position of each edge portion using the table position information and the signal of the photodetector device (42), and computing the coordinates of a mark center (P) based on the data.

7. A method according to claim 6, characterized in that said table structure (10) is moved so that the photodetector device (42) crosses the first edge portion, a plurality of first multi-level scanning signals (46) being produced in response to a light component (36) at the time of that movement of said table (10); and in that said table (10) is moved reversely so that the photodetector device (42) crosses the second edge portion, a plurality of second multi-level scanning signals (46) being produced in response to the light component (36) at the time of that movement of said table (10); edge data being produced in response to the first and second scanning signals.

8. A method according to claim 7, characterized in that said table (10) is moved so that the photodetector device (42) crosses the third edge portion, a plurality of third multi-level scanning signals (46) being produced in response to the light component (36) at the time of that movement of said table (10); and in that said table (10) is moved so that the photodetector device (42) crosses the fourth edge portion, a plurality of fourth multi-level scanning signals (46) being in response to the light component (36) at the time of that movement of said table (10); edge data being produced in response to the third and fourth scanning signals.

9. A method according to claim 6, characterized in that said photodetector device (42) includes a plurality of photosensing elements (44), in that there is generated from the signal of each of said photosensing element (44) a multi-level scanning signal (46) quantitized at predetermined time intervals, and in that said signal serves to compute one of the edge data.

10. A method according to claim 9, characterized in that a plurality of scanning signals (46) with signal levels varying with time according to the light component (36) which is generated when said photodetector device (42) crosses one mark edge portion during the movement of the table structure is supplied from said photosensor elements (44), and in that the envelope waveform (54, 56) of the signal levels of each of said scanning signals (46) is computed to compute the position of said mark edge portion at a detection error smaller than the pitch in the array of said photosensor elements (44) on the basis of a combination of each of said envelope waveforms (54,56).

11. A method according to claim 10, characterized in that said photosensor elements (44) are linearly arrayed in one of the table movement directions.

12. A method according to claim 10, characterized in that said light component contains the light transmitted through or reflected from said workpiece (12).